R E S E A R C H A R T I C L E
Open Access
Value of reduced glomerular filtration rate
assessment with cardiometabolic index:
insights from a population-based Chinese
cohort
Hao-Yu Wang
1†, Wen-Rui Shi
1†, Xin Yi
2, Shu-Ze Wang
3, Si-Yuan Luan
4and Ying-Xian Sun
1*Abstract
Background:Recent studies have suggested that cardiometabolic index (CMI), a novel estimate of visceral adipose tissue, could be of use in the evaluation of cardiovascular risk factors. However, the potential utility and clinical significance of CMI in the detection of reduced estimated glomerular filtration rate (eGFR) remains uncertain. The purpose of this study was to investigate the usefulness of CMI in assessing reduced eGFR in the general Chinese population.
Methods:This cross-sectional analysis included 11,578 participants (mean age: 53.8 years, 53.7% females) from Northeast China Rural Cardiovascular Health Study (NCRCHS) of general Chinese population (data collected from January 2013 to August 2013). CMI was calculated by triglyceride to high density lipoprotein cholesterol ratio multiply waist-to-height ratio. Reduced eGFR was defined as eGFR< 60 ml/min per 1.73m2. Multivariate regressions were performed to determine CMI’s association with eGFR value and eGFR reduction, ROC analyses were employed to investigate CMI’s discriminating ability for decreased eGFR.
Results:The prevalence of reduced eGFR was 1.7% in males and 2.5% in females. CMI was notably more adverse in
reduced eGFR groups, regardless of genders. In fully adjusted multivariate linear models, each 1 SD increment of CMI caused 3.150 ml/min per 1.73m2and 2.411 ml/min per 1.73m2loss of eGFR before CMI reached 1.210 and 1.520 in males and females, respectively. In logistic regression analyses, per 1 SD increase of CMI brought 51.6% additional risk of reduced eGFR in males while caused 1.347 times of risk in females. After divided into quartiles, people in the top quartile of CMI had higher adjusted ORs of having reduced eGFR, with ORs of 4.227 (1.681, 10. 627) and 3.442 (1.685–7.031) for males and females respectively. AUC of CMI was revealed to be 0.633 (0.620–0.646) in males and 0.684 (0.672–0.695) in females.
Conclusions:Higher CMI was independently associated with greater burden of reduced eGFR, highlighting VAT
distribution and dysfunction as a potential mechanism underlying the association of obesity with kidney damage and adverse cardiovascular outcomes. The findings from this study provided important insights regarding the potential usefulness and clinical relevance of CMI in the detection of reduced eGFR among general Chinese population.
Keywords:Cardiometabolic index, Dyslipidemia, Obesity, Reduced eGFR, Sex-specific, Visceral adipose tissue
* Correspondence:[email protected]
†Hao-Yu Wang and Wen-Rui Shi contributed equally to this work.
1Department of Cardiology, The First Hospital of China Medical University,
155 Nanjing North Street, Heping District, Shenyang 110001, China Full list of author information is available at the end of the article
Background
Chronic kidney disease (CKD) was a disease with heterogeneous etiology that caused 15.8 deaths per 100,000 people worldwide in 2013 and was estimated to attack 119.5 million Chinese people in the year of 2012 [1, 2]. Besides, CKD has been identified to be related with multiple cardiovascular risk factors like hyperten-sion and diabetes mellitus (DM), revealing its close as-sociation with cardiovascular health [3, 4]. Therefore, CKD has brought great burden to the health-care work and economy [5, 6]. Knowing the poor outcomes and high cost that caused by CKD as a worldwide public health problem, it is a good choice to focus on the early detection of reduced estimated glomerular filtration rate (eGFR) for avoiding the progression of CKD. Ac-cordingly, early detection of reduced eGFR is critically needed for timely CKD prevention and overall improve-ment in prognosis.
There is an wealthy evidence that dyslipidemia and obesity are strongly related to the deterioration of renal function, even in the early stage [7, 8]. Dyslipidemia in CKD often exhibits a pattern of increased triglyceride (TG) levels and decreased high density lipoprotein cholesterol (HDL-C) levels [7]. Findings revealed that increased TG level was significantly correlated with CKD [9–11]. Furthermore, as a combination of 2 char-acteristics of dyslipidemia in CKD, TG to HDL-C ratio (TG / HDL-C) has also been identified to strongly asso-ciated with decline of eGFR in participants without CKD and rapid decrement of eGFR in participants with CKD [12–14]. Obesity is another condition that strongly correlated with decreased eGFR or CKD [8, 15]. As the most widely used index of obesity, body mass index (BMI) was revealed to correlated with reduced eGFR, CKD and end stage renal disease (ESRD) [16,17]. How-ever, BMI did not take abdominal obesity into consider-ation when reflecting obesity status. Therefore, some
indexes such as waist circumference (WC) and
waist-to-height ratio (WHtR) were designed to be mea-surements of central obesity [18]. And they have already showed strong associations with CKD [19–22]. Nevertheless, comparisons between BMI, WC and WHtR did not provide sufficient evidence for anyone of them to be a premier marker of CKD with great sensi-tivity or specificity [19–22]. Moreover, indexes of ab-dominal obesity were insufficient to detecting visceral adipose tissue (VAT), which was elucidated to possess a more adverse effect on CKD development than sub-cutaneous adipose tissue (SAT) [23,24]. However, when cooperating with aforementioned indexes of dyslipidemia, their strength to discriminating VAT got reinforcement [24]. Taken together, combination of dyslipidemia and ab-dominal obesity could improve the identification of VAT and therefore refine the detecting of reduced eGFR.
Recently, a novel marker named“cardiometabolic index (CMI)” has been put forward by Ichiro Wakabayashi [25]. CMI can be considered as an ideal marker to recognize VAT because of its integration of dyslipid-emia and abdominal obesity. This emerging marker has been shown in several studies to be a useful screening tool for various populations, identifying those with a deteriorated metabolic profile and higher risk for car-diovascular disease, such as left ventricular geometry abnormalities, hyperuricemia, diabetes, hypertension, and ischemic stroke [25–29]. It is unclear, however, if CMI can be an identifier of reduced eGFR, independent of cardiovascular risk factors and hypertension. Accord-ingly, the present study was designed to test whether or not higher CMI increase the risk of reduced eGFR in the general Chinese population.
Methods
Study population
This study was part of a large cross-sectional population-based epidemiological investigation that described the prevalence, incidence, and natural history of cardiovascu-lar risk factors among 11,956 permanent residents (≥35 years of age) in rural areas of China from January 2012 to August 2013. The full details regarding the design and rationale of the study were extensively described else-where [26,30,31]. Briefly, the study adopted a multistage, stratified random cluster-sampling scheme. In the first stage, 3 counties (Dawa, Zhangwu, and Liaoyang County) were selected from the eastern, southern, and northern region of Liaoning province. In the second stage, 1 town was randomly selected from each county (a total of 3 towns). In the third stage, 8 to 10 rural villages from each town were randomly selected (a total of 26 rural villages). Participants with pregnancy, malignant tumor, or mental disorder were excluded from the present study. The study protocol complied with the Second Helsinki Declaration (and recent amendments) and was approved by the Ethics Committee of China Medical University (Shenyang, China), and all study participants provided written in-formed consent. Patients with missing data for any of CMI components or other variables analyzed in the study (n= 378) were excluded. Accordingly, the final study co-hort was composed of 11,578 subjects.
Data collection and measurements
medicine usage condition, history of cardiovascular disease (CVD).
A carefully designed questionnaire was used to collect data from subjects. A central steering committee with a subcommittee conducted the quality control of the infor-mation collection process. Subjects were divided into two groups: Han and others. Education level was regarded as three ordinal groups: primary school or below, middle school, high school or above. Three ordinal groups (≤5000 CNY,5000–20,000 CNY,> 20,000 CNY) were utilized to represent the family annual income level. Diet score was calculated according to the collected information about vegetable and meat consumption, a lower score meant lower meat consumption and higher vegetable consump-tion. Detailed information about diet score has been re-ported in a prior study from our team [32]. Physical activity level was acquired in both work and off hours, and divided into three levels (low, middle, high). Messages of current smoking and alcohol intake status were col-lected through a series of questions about smoking and drinking history of participants. Medicine usage such as anti-hypertensive, anti-diabetic and lipid-lowering drugs was recorded in the questionnaire as well. History of CVD included coronary heart disease, arrhythmia, heart failure, and stroke. Meanwhile, history of kidney disease was de-fined as nephritis, kidney dysfunction, renal stones, renal tumor, and autoimmune kidney disease.
After subjects completed a 5 min rest in sitting position with relaxation, two randomly selected trained medical staffs performed the blood pressure measurements for them. Three consecutive readings were recorded and their mean value was taken into statistical analyses.
With regard to the measurements of anthropometric indices, subjects were requested to wear in light clothing without shoes. Standard weight was measured to the nearest 0.1 kg by using a calibrated digital scale. Subjects held in a standing position when a calibrated stadi-ometer was used to quantify their standard height to the nearest 0.1 cm. As for the measurement of WC, elastic measuring tapes was used to get the readings in a hori-zontal position at 1 cm above the umbilicus. All of the above measurements were performed twice and their mean values were used for analyses.
Detailed delineation of the process of storage and methods of laboratory measurements was reported in our previous studies [26, 30, 31]. Briefly, fasting (12 h overnight) blood samples were collected through venipuncture. And these samples were separated and frozen at −20 °C within 1 h after the collection, and then transported to a laboratory with certification for examination. An Olympus AU 640 auto analyzer (Olympus, Kobe, Japan) was used for biochemical analyses. All labora-tory equipment was calibrated and blinded duplicate sam-ples were used.
Definitions
BMI was calculated as mean weight divided by mean height squared (kg/m2). WHtR was acquired as WC di-vided by mean height. CMI was obtained by the following equation [25]: CMI = TG/HDL-C × WHtR. lipid accumu-lation product (LAP) was calculated according to the sex-specific formula [33]: LAP = TG (mmol/L) × [WC (cm)-58] for females and LAP = TG (mmol/L) × [WC (cm)-65] for males. And visceral adiposity index (VAI) was determined by using the following formula [34]: Males: VAI = [WC / 39.68 + (1.88 × BMI)] × (TG / 1.03) × (1.31 / HDL); Females: VAI = [WC / 36.58 + (1.89 × BMI)] × (TG / 0.81) × (1.52/HDL).
eGFR was calculated according to CKD-EPI equation [35]. Reduced eGFR was defined as eGFR < 60 ml/min per 1.73m2[36]. Diagnostic criteria of hypertension were mean systolic blood pressure (SBP) equal to or greater than 140 mmHg and/or diastolic blood pressure (DBP) at least 90 mmHg and/or participants who were on anti-hypertensive medications or self-reported previous diagnosed hypertension [37]. Diagnosis of diabetes based on the American Diabetes Association criteria: fasting plasma glucose (FPG)≥7.0 mmol/L and/or self-reported previous diagnosis history or receiving plasma glucose lowering therapy [38].
Statistical analyses
the results were displayed as ORs and 95% confidence intervals (95% CI). Lastly, receiver-operating character-istic (ROC) curve was employed to investigate the opti-mal cut-off value of CMI to detect the presence of reduced eGFR. Area under the curve (AUC) was used to compare the discriminating ability of CMI and other indexes. All of the statistical analyses involved were performed by SPSS 25.0 software (IBM corp), statistical software packages R (http://www.R-project.org, The R Foundation) and EmpowerStats ( http://www.empower-stats.com, X&Y Solutions, Inc., Boston, MA), Prism 7.0 software (Graphpad software, Inc) and MedCalc version 18.5 (MedCalc software, Belgium), a two-tailedP value less than 0.05 indicated statistical significant.
Results
After excluding ineligible participants, we finally took 11,578 subjects into analyses, the results were shown in Table1. Of these subjects, 46.3% were males. The mean age of total population was 53.8 years while people with eGFR decrement were elder than their counterparts in both genders. The prevalence of reduced eGFR was 1.7% in males while females exhibited a higher rate of 2.5%. As for the demographic information, education level, family annual income, diet score and physical activity ex-hibited significant lower levels in patients with reduced eGFR. Contradictorily, female patients were more likely to smoke at present but male patients tended not to be a current smoker. Only males showed significant differ-ence on alcoholic status between groups, but the results revealed that normal eGFR subjects had greater possibil-ity to be a current drinker. With regard to laboratory test results, FPG, serum uric acid, TG and TG/HDL-C confirmed a remarkable augmentation among eGFR dec-rement patients, together with a dramatically decline of HDL-C in patients. SBP, height, WC, WHtR showed great increment when comparing CKD patients with their counterparts regardless of sex. But DBP only exhib-ited higher level in male patients while less weight was specific to female patients. Prevalence of hypertension and diabetes were noticeable higher among subjects with reduced eGFR for both genders, and the usages of anti-hypertensive, anti-diabetic and lipid-lowering drug were more popular in reduced eGFR group. Concordantly, the prevalence of cardiovascular and kidney diseases his-tory showed dramatic augmentation in eGFR reduction patients of both genders. Within our expectation, patients with reduced eGFR of both genders experienced pro-nounced increment of CMI, VAI as well as LAP.
Quartile analyses revealed gradient correlation between CMI and prevalence of reduced eGFR, as displayed in Fig.1. When comparing the top quartile with the bottom one, males exhibited a 6.0-fold change for the probability of eGFR reduction while females showed a higher fold
change of 8.2. Both genders confirmed linear trends of prevalent eGFR reduction across quartiles of CMI (all P for trend< 0.001).
Results of linear regression analyses identified robust negative association between CMI with eGFR. As presented in Table 2, logarithmic likelihood ratio test identified the association between CMI and eGFR was non-linear in both genders after adjustment of age, race, education level, family annual income, diet score, phys-ical activity, current smoking, alcohol intake, hyperten-sion, diabetes, antihypertensive drug, antidiabetic drug, lipid-lowering drug, history of cardiovascular and kidney diseases. In the full model of males, the two-piecewise linear model revealed a rapid decrease of eGFR value along with the early increase of normalized CMI, with a βvalue of−3.150 (−3.589,−2.712). However, when nor-malized CMI reached 1.113 (equaled to 1.210 of CMI), the change of eGFR shifted to a mild increase, with aβ value of 1.906 (0.675, 3.138). Similarly, in the complete model of females, the eGFR also suffered a loss of 2.411 ml/min per 1.73m2for each SD increase of CMI before normalized CMI arrived to 1.472 (equaled to 1.520 of CMI). But after that, the eGFR began to have a increment of 2.268 ml/min per 1.73m2along with every single unit increase of normalized CMI.
Logistic regression analyses confirmed the intensive relationship between CMI with reduced eGFR, the re-sults were arranged in Table 3. In males, per 1 SD in-crease of CMI still caused 52% of additional risk for eGFR reduction after adjusting all included cofounders. When divided into quartiles, highest quartile of CMI displayed 4.2 times risk for developing eGFR decrement compared with lowest category in the full model, and there was a significant linear trend for the risk of eGFR reduction across the quartiles of CMI (P for trend< 0.001). Similar results were observed in female participants, risk of eGFR reduction got an elevation of 34.7% for every single SD increment of CMI. Furthermore, the ORs for top quartile compared with bottom group was 3.442 (1.685–7.031), and the linear trend across quar-tiles also existed (P for trend< 0.001). We further conducted a test for the interaction between genders and CMI, and the results displayed to be insignificant, identifying the robust association between CMI and re-duced eGFR across genders.
ROC analyses showed a significant AUC value of CMI for discriminating eGFR reduction, the results were summarized in Table 4. In males, our findings displayed that CMI possessed the greatest AUC (AUC: 0.633, 95% CI: 0.620–0.646) among various indexes, statistically sig-nificantly higher than that of LAP (AUC 0.633 vs. 0.606,
Table 1Characteristics of participants with reduced eGFR stratified by sex
Variables Males (n= 5360) Females (n= 6218)
Reduced eGFR (n= 90) Normal eGFR (n= 5270) Pvalue* Reduced eGFR (n= 155) Normal eGFR (n= 6063) Pvalue*
Age (years) 68.78 ± 10.41 54.11 ± 10.64 < .001 68.69 ± 8.87 52.99 ± 10.08 < .001
Race (Han) (%) 87 (96.7) 4987 (94.6) 0.394 152 (98.1) 5747 (94.8) 0.068
Education level (%) < .001 < .001
Primary school or below 57 (63.3) 2177 (41.3) 135 (87.1) 3398 (56.0)
Middle school 26 (28.9) 2487 (47.2) 18 (11.6) 2187 (36.1)
High school or above 7 (7.8) 606 (11.5) 2 (1.3) 478 (7.9)
Income level (CNY) (%) < .001 < .001
≤5000 28 (31.1) 693 (13.1) 41 (26.5) 680 (11.2)
5000–20,000 45 (50.0) 2827 (53.6) 77 (49.7) 3361 (55.4)
> 20,000 17 (18.9) 1750 (33.2) 37 (23.9) 2022 (33.3)
Diet score 2.24 ± 0.96 2.55 ± 1.10 0.008 1.68 ± 1.03 2.15 ± 1.11 < .001
Physical activity (%) < .001 < .001
Low 52 (57.8) 1158 (22.0) 100 (64.5) 2133 (35.2)
Middle 36 (40.0) 3817 (72.4) 42 (27.1) 3587 (59.2)
High 2 (2.2) 295 (5.6) 13 (8.4) 343 (5.7)
Current smoking (%) 28 (31.1) 3030 (57.5) < .001 43 (27.7) 987 (16.3) < .001
Current alcohol intake (%) 14 (15.6) 2421 (45.9) < .001 3 (1.9) 180 (3) 0.452
FPG (mmol/L) 5.74 (5.35–6.67) 5.60 (5.22–6.09) 0.010 5.78(5.30–7.05) 5.49 (5.12–5.99) < .001
Serum uric acid (μmol/L) 447.52 ± 122.56 331.80 ± 81.34 < .001 363.21 ± 94.01 253.06 ± 64.66 < .001
TG (mmol/L) 1.41 (1.10–2.07) 1.22 (0.86–1.88) 0.006 1.86 (1.30–2.55) 1.24 (0.89–1.88) < .001
HDL-C (mmol/L) 1.21 ± 0.27 1.41 ± 0.42 < .001 1.35 ± 0.41 1.41 ± 0.34 0.018
TG/HDL-C 1.19 (0.80–2.11) 0.91 (0.57–1.57) < .001 1.42 (0.92–2.24) 0.92 (0.59–1.51) < .001
SBP (mmHg) 161.58 ± 26.32 143.36 ± 22.43 < .001 154.16 ± 26.14 139.79 ± 23.87 < .001
DBP (mmHg) 89.01 ± 16.20 83.69 ± 11.71 0.003 82.57 ± 14.77 80.52 ± 11.41 0.088
Height (cm) 164.08 ± 5.56 166.46 ± 6.35 < .001 152.91 ± 5.91 155.69 ± 6.06 < .001
Weight (kg) 68.52 ± 12.06 68.61 ± 11.10 0.940 58.08 ± 10.64 60.33 ± 10.11 0.006
BMI (kg/m2) 25.36 ± 3.65 24.72 ± 3.54 0.091 24.77 ± 3.95 24.86 ± 3.76 0.778
WC (cm) 87.05 ± 10.36 83.74 ± 9.75 0.003 84.07 ± 10.37 81.20 ± 9.71 < .001
WHtR 0.53 ± 0.06 0.50 ± 0.06 < .001 0.55 ± 0.07 0.52 ± 0.06 < .001
hypertension (%) 74 (82.2) 2818 (53.5) < .001 120 (77.4) 2907 (47.9) < .001
diabetes (%) 22 (24.4) 508 (9.6) < .001 41 (26.5) 629 (10.4) < .001
Anti-hypertensive drug (%) 50 (55.6) 641 (12.2) < .001 73 (47.1) 989 (16.3) < .001
Anti-diabetic drug (%) 13 (14.4) 148 (2.5) < .001 20 (12.9) 278 (4.6) < .001
Lipid-lowering drug (%) 7 (7.8) 159 (3.0) 0.021 12 (7.7) 202 (3.3) 0.003
history of CVD (%) 46 (51.1) 781 (14.8) < .001 62 (40.0) 1092 (18.0) < .001
history of kidney disease 5 (5.6) 60 (1.1) < .001 11 (7.1) 77 (1.3) < .001
eGFR (ml/min per 1.73m2) 54.81 (46.46–57.67) 95.44 (87.00–103.44) < .001 52.90 (46.89–57.47) 93.23 (82.89–103.29) < .001
LAP (cm·mmol/L) 31.88 (18.63–51.35) 21.79 (10.62–43.35) 0.001 44.94 (26.00–77.56) 27.84 (16.00–50.25) < .001
VAI 1.51 (1.00–2.60) 1.12 (0.68–1.98) < .001 2.68 (1.78–4.17) 1.67 (1.06–2.78) < .001
CMI 0.62 (0.41–1.10) 0.45 (0.27–0.82) < .001 0.81 (0.49–1.26) 0.47 (0.29–0.81) < .001
Data are expressed as mean ± standard deviation(SD) or median (interquartile range) and numbers (percentage) as appropriate. Income level: family annual income;CNYChinese currency (1CNY = 0.15 USD),SBPsystolic blood pressure,DBPdiastolic blood pressure,FPGfasting plasma glucose,CVDcardiovascular disease, includes coronary heart disease, arrhythmia, heart failure, and stroke; Kidney disease: includes nephritis, kidney dysfunction, renal stones, renal tumor, and autoimmune kidney disease. Body mass index;WCwaist circumference,WHtRwaist-to-height ratio,TGtriglyceride,HDL-Chigh density lipoprotein cholesterol,TG/HDL-Ctriglyceride to high density lipoprotein cholesterol ratio,eGFRestimated glomerular infiltration rate,LAPlipid accumulation product, VAIvisceral adiposity index,CMIcardiometabolic index
sensitivity of 88.9% and a specificity of 37.1%. In fe-males, CMI had an AUC (AUC: 0.684, 95% CI: 0.672– 0.695) which was comparable to that of VAI (AUC: 0.688, 95% CI: 0.676–0.699, P value for difference = 0.148) and statistically significantly greater than that of LAP (AUC 0.684 vs 0.660, P= 0.047), WC (AUC 0.684 vs 0.588,
P< 0.001) and BMI (AUC 0.684 vs 0.500,P< 0.001). Fur-thermore, CMI exhibited the greatest sensitivity (89.0%) in this gender. However, the specificity was still low and was given to 39.8%.
Discussion
Our study showed that in this large, community-based cohort of middle-aged Chinese individuals, CMI, a novel measure of VAT, was independently and robustly
associated with the presence of reduced eGFR in both genders. CMI could reflect not only VAT but also patho-logic process that resulted in impaired kidney function, thus provided a clinical clue for related basic researches. As our results revealed the potential of CMI as a screen-ing marker of reduced eGFR, further understandscreen-ing of the underlying pathophysiology should improve strategies to prevent and potentially reverse detrimental kidney failure and cardiovascular outcomes, especially in people with poor socioeconomic conditions.
Dyslipidemia is a condition that always appears in the development of CKD, even in the early stage of eGFR reduction, and the major components of dyslipidemia in CKD has been revealed to be increased TG level and decreased HDL-C level [7]. Indexes that represent this specific pattern has been identified to have a strong correl-ation with CKD. For instance, Hou et al. revealed that TG level was closely associated with mildly decreased eGFR even among middle aged or elderly Chinese population. It was worthy to note that this relationship existed when TG level was still in the normal range. This finding gave us an implication that increased TG level could appear in the early process of decreased eGFR [9]. Lee et al. identified hypertriglyceridemia as an independent risk factor of pro-gressed CKD, confirmed the association persistent in mid-dle and late stage of CKD [11]. Furthermore, Tozawa et al. and Shimizu et al. provided compelling evidences that TG had an independent effect to increase the risk of eGFR decrement from longitudinal studies, Tozawa et al. also identified HDL-C positively related to eGFR value and re-vealed TG had an impact on the development of protein-uria [10,40]. Apart from above investigations, researches also identified the utility of TG/HDL-C in identifying eGFR reduction as it incorporated both TG and HDL-C level. Ho et al. demonstrated the correlation between TG/ HDL-C with CKD, and Wen et al. even showed the advan-tage of TG/HDL-C when compared with TG alone with respect to the development of CKD, confirmed our afore-mentioned hypothesis [13, 14]. However, there was one point that worth to be mentioned, although TG/HDL-C had a robust association with CKD, it was not an ideal marker of eGFR reduction due to its low sensitivity and specificity [13].
As another condition that always accompany with CKD, obesity is related to multiple risk factors of CKD, such as hypertension and diabetes [41, 42]. Its direct association with CKD has been evaluated through in-dexes. As the most widely used index of obesity, BMI has already been elucidated to have a connection with new-onset CKD, CKD progression and end stage renal failure [16, 17, 43]. However, since BMI do not consider adipose tissue distribution, there is no doubt that BMI is unsuitable for reflecting the relationship between obesity phenotype with CKD. Therefore, indexes of abdominal Fig. 1The prevalence of reduced eGFR by quartiles of CMI.
Prevalent decreased eGFR increased proportionally across ascending quartiles of CMI in both genders (P for trend< 0.05). Abbreviations: CMI, cardiometabolic index; eGFR: estimated glomerular
obesity such as WC and WHtR have been widely investi-gated for their associations with CKD. Among them, WHtR was found to have greatest application potential for detecting CKD by several studies [19,20]. Theoretic-ally, by using height to standardize WC value, WHtR was design to be a superior reflection of abdominal obesity since people with different height were supposed to have different standard WC value. Nevertheless, stud-ies did not support any of these indexes to be an eligible marker with great sensitivity or specificity for detection of reduced eGFR [19–22]. Meanwhile, studies found that VAT contributed more to CKD than SAT [23]. Indexes like WC hadd no power to distinguish VAT from SAT alone but could recognize VAT correctly when working
together with TG [24,44]. Taking all above messages to-gether, we can speculate that indexes contain informa-tion about dyslipidemia and abdominal obesity should do better in the identification of VAI and be able to re-fine the recognition of eGFR reduction.
Indexes that integrate both abdominal obesity and dys-lipidemia has been proposed. Determined by WC and TG levels, LAP was put forward for recognizing cardiovascu-lar risk [33], and it has also been identified to possess a stronger association with CKD than BMI, WC and WHtR [45]. Later, a concept named VAI was posited as an indica-tor of cardiometabolic risk with an additional consider-ation of HDL-C and BMI when compared with LAP [34]. In recent studies, VAI showed strong association with
Table 2Evaluation of the impact of CMI on eGFR value with the use of piecewise linear regressiona
Unadjusted MV adjusted
Males
Linear model βValue (95% CI)Pvalue −1.850 (−2.259,−1.440) < 0.001 −2.214 (−2.561,−1.866) < 0.001
Non-linear model Breakpoint (K) 1.024 1.113
β1 (<K) (95% CI)Pvalue −3.324 (−3.863,−2.786) < 0.001 −3.150 (−3.589,−2.712) < 0.001
β2 (>K) (95% CI)Pvalue 3.777 (2.371, 5.182) < 0.001 1.906 (0.675, 3.138) 0.002 Logarithmic likelihood ratio testPvalue < 0.001 < 0.001
Females
Linear model βValue (95% CI)Pvalue −4.046 (−4.438,−3.654) < 0.001 −1.914 (−2.262,−1.566) < 0.001
Non-linear model Breakpoint (K) 1.34 1.472
β1 (<K) (95% CI)Pvalue −5.001 (−5.472,−4.530) < 0.001 −2.411 (−2.813,−2.010) < 0.001
β2 (>K) (95% CI)Pvalue 2.478 (0.636, 4.320) 0.008 2.268 (0.547, 3.988) 0.010 Logarithmic likelihood ratio testPvalue < 0.001 < 0.001
Abbreviations:CMIcardiometabolic index,ORodds ratio,95% CI95% confidence interval. Linear model: model that presumes the association between CMI and eGFR is linear. Non-linear model: model that presumes the association between CMI and eGFR is non-linear and has breakpoint. Unadjusted: no adjustment; MV adjusted: multivariable adjusted model, includes age, race, education level, family annual income, physical activity, current smoking, current alcohol intake, hypertension, diabetes, antihypertensive drug, antidiabetic drug, lipid-lowering drug, history of cardiovascular disease and kidney disease.aTwo-step linear regression model was applied to explore the non-linear association between CMI and eGFR
Table 3Sex-specific logistic regression models for reduced eGFR with CMI
Males Females Pfor
interaction Unadjusted OR
(95% CI) P
value MV adjusted OR
(95% CI) P
value Unadjusted OR
(95% CI) P
value MV adjusted OR
(95% CI) P
value
CMI (Per 1 SD increase)
1.467 (1.211, 1.777) < 0.001 1.516 (1.182, 1.943) 0.001 1.760 (1.519, 2.040) < 0.001 1.347 (1.122, 1.618) 0.001 0.172
Quartiles of CMI
Quartile 1 1.000 1.000 1.000 1.000
Quartile 2 3.711 (1.500, 9.182) 0.005 2.717 (1.061, 6.957) 0.037 2.835 (1.372, 5.856) 0.005 2.285 (1.069, 4.888) 0.033
Quartile 3 5.092 (2.112, 12.273) < 0.001 3.867 (1.551, 9.640) 0.004 4.187 (2.090, 8.387) < 0.001 2.509 (1.204, 5.225) 0.014
Quartile 4 5.439 (2.267, 13.052) < 0.001 4.227 (1.681, 10.627) 0.002 7.939 (4.091, 15.408) < 0.001 3.442 (1.685, 7.031) < 0.001 Pvalue for
trend
< 0.001 0.001 < 0.001 < 0.001
CKD and greater potential as a marker of eGFR reduction than LAP as well as traditional anthropometric indexes [45,46]. However, these two novel indexes still have limi-tations. Both LAP and VAI have sex-specific equations, and VAI has a sophisticated algorithm, these will increase the complexity of calculation. As for LAP, information about HDL-C is not utilized, which is an important aspect of dyslipidemia in CKD. Moreover, height info is also wasted by LAP, but people with different height should have different WC standard. Last but not least, although VAI possessed a greater discriminative power than LAP and other indexes, it still failed to reach satisfying sensitiv-ity and specificsensitiv-ity values, thus its value for becoming a marker of eGFR reduction is limited [46].
CMI is a young index which was posited by Ichiro Wakabayashi in 2015 [25]. Different from LAP and VAI, it reflects both characteristics of dyslipidemia in CKD and accurate central obesity status through a simple, general-ized eq. CMI has also been identified to correlate with multiple cardiovascular risk factors, implicated to have a great potential in screening related diseases [25–29]. Since many of these diseases has been identified to be risk fac-tors of CKD [16,47–50], we hypothesize that CMI corre-lates with reduced eGFR and is capable of being a premier indicator of eGFR decrement. In our present study, re-gression analyses revealed the intensive and direct associa-tions of CMI with eGFR value and eGFR reduction in both genders, confirming above hypothesis. Moreover, we observed a breakpoint in the association between CMI and eGFR value, which meant the change of eGFR shifted from decrease to increase after CMI reached certain value. And this phenomenon was robust and persistent across genders. Although the exact mechanism underlying this
paradoxical phenomenon is still unclear, some previous findings can also give us a clue for possible explanations. While obesity was commonly considered to associate with the decline of eGFR [8,15], studies also found a protective effect of obesity on CKD patients [51, 52], and this pro-tective effect could be explained by the evidence that ex-cessive adipose tissue might temper the deleterious effects of inflammatory factors by sequestering them [53, 54]. Consistent with this theory, the breakpoint was located at high levels of CMI (> 75 percentile) in each gender, which meant that the protective effect became more predomin-ant than the destructive effect only when fatty tissue accumulated to a sufficient volume. Another possible ex-planation was the residual confounders. Although we had adjusted several factors that related to CMI and eGFR, there were still some confounders that we did not take into our regression equations. Therefore, some uncovered confounding factors could influence our results and caused this paradox. In order to validate the paradox founded in the present study, studies with prospective de-sign and consideration of more modifiers are needed in the future.
Further, we confirmed the association between CMI and reduced eGFR was positive and strong. We found that the risk of reduced eGFR was proportionally in-creased across the quartiles of CMI, which was different from the breakpoint phenomenon in the association be-tween CMI and eGFR value. Given that the breakpoint of CMI located behind the 75 percentile in both genders, we could easily figure out that the limited increase of eGFR in CMI levels above the breakpoint has minimal effect on the whole inverse trend of CMI with reduced kidney function (eGFR< 60 ml/min per 1.73m2). We also
Table 4AUC for indexes to discriminate eGFR reduction in females and males
Variables AUC (95%CI) Pvalue Cut-off according to Youden’s index Sensitivity (%) Specificity (%)
Males
CMI 0.633 (0.620–0.646)a,c < 0.001 > 0.35 88.9% 37.1%
VAI 0.627 (0.614–0.640)c < 0.001 > 0.83 90.0% 34.7%
LAP 0.606 (0.593–0.619)c < 0.001 > 18.33 77.8% 43.6%
WC 0.589 (0.576–0.602)c 0.001 > 79.1 83.3% 35.0%
BMI 0.544 (0.531–0.558) 0.135 > 22.02 88.9% 22.8%
Females
CMI 0.684 (0.672–0.695)a,b,c < 0.001 > 0.39 89.0% 39.8%
VAI 0.688 (0.676–0.699)a,b,c < 0.001 > 1.69 80.0% 50.4%
LAP 0.660 (0.648–0.672)b,c < 0.001 > 36.12 65.2% 61.8%
WC 0.588 (0.576–0.600)c < 0.001 > 84.20 52.9% 64.0%
BMI 0.500 (0.488–0.513) 0.995 ≤27.81 76.1% 20.1%
Abbreviations:AUCarea under the ROC curve,95% CI95% confidence interval,CMIcardiometabolic index,VAIvisceral adiposity index,LAPlipid accumulation product,WCwaist circumference,BMIbody mass index
a
indicates a significant larger as compared to LAP; b
indicates a significant larger as compared to WC; c
noticed that males had a slightly higher risk of develop-ing reduced eGFR when compared with their female counterparts. But the interaction test revealed that this disadvantage was statistically insignificant. However, al-though we could not conclude that males were under greater risk of reduced eGFR, we still elucidated that CMI had a robust association with eGFR reduction in both genders, identifying the utility and stability of CMI for the risk stratification of reduced eGFR.
Findings from ROC analyses provided compelling evi-dence for CMI to be a screening marker of CKD. In males, CMI possessed greatest AUC value, although it did not have a significant larger discriminating ability than VAI, the simple and generalized equation for both sexes would be a vantage. With a sensitivity of 88.9%, CMI demonstrated its capability to function well in the screening test, since more suspicious patients should be positive in this kind of examination. It was worthy to note that neither of CMI, VAI and LAP had a significant greater AUC than WC, which was beyond our expect-ation, this could be explained by the lacking of sample size since the absolute number and prevalence of re-duced eGFR were much lower in males than in females. Further studies are needed to validate this intriguing finding. In females, CMI did an even better job as it per-formed better than other indexes significantly except VAI. Although exhibited a negligible disadvantage in dis-criminating eGFR reduction, CMI showed a much higher sensitivity than VAI, thus the applicative value for CMI was still superior than VAI among females. In con-clusion, with great discriminative ability and high sensi-tivity value, CMI showed its potential for becoming an economic screening marker to filter out people with re-duced eGFR.
Mounting evidence has emerged to reveal the mech-anism under the association between CMI with CKD. First, mechanism of dyslipidemia in CKD has already been elucidated. Hypertriglyceridemia is a multifactorial phenomenon, and is partially due to diminished catab-olism. This reduction is an outcome of depressed lipo-protein lipase (LPL) activity, which is responsible for the hydrolysis of TG [55]. And the decreased LPL activ-ity has been identified to be caused by secondary hyper-parathyroidism induced insulin resistance (IR) and excess of lipase inhibitors like apolipoprotein (Apo) C-III in the progression of CKD [56, 57]. For another aspect, hepatic Apo A-I gene expression and hepatic lecithin cholesterol acyl transferase (LCAT) mRNA ex-pression are downregulated during the course of CKD, and then this change attributes to a lower plasma Apo A-I level and decreased LCAT activity [58, 59]. Since Apo A-I is an essential functional component of HDL-C [60], the concentration of HDL-C in blood is consequently decreased. Similarly, the normal function
of HDL-C is also impaired because of inadequate LCAT activity, which is important in the process of transport-ing cholesterol to HDL in peripheral tissues [60]. Second, basic research has also revealed the intrinsic mechanism between obesity with CKD. Adipocytes produce a series of factors, such as angiotensinogen, precursor of angiotensin II and leptin, which may trigger glomerular hypertension through local renin-angiotensin-aldosterone system (RAAS) or sympathetic system, then leading to the decrease of eGFR [8]. Further-more, inflammation involves in the association between obesity with CKD as well, macrophages are found to accu-mulate in the kidney of obese animals in the effect of angiotensin [61]. Some inflammatory cytokines such as interleukin-6 (IL-6) and Tumor necrosis factor-α(TNF-α) have also been implicated in obesity-related CKD [62]. Therefore, obesity could promote the progression of renal disease through several, partially overlapping mechanisms. Sum up above two points, as composites of CMI, dyslipid-emia and obesity have experiment confirmed relationship with CKD. Thus, the association between CMI with re-duced eGFR observed in our study has a concrete founda-tion from basic researches.
There are still some limitations in our study that needed to be mentioned. First, due to the cross-sectional design, our work can only provide evidence about the strong association between CMI and reduced eGFR, but information about the causality of this relationship need further prospective studies to confirm. Second, it has an inherent limitation of not being a randomized study. Al-though multivariable adjustments were performed for potential confounding factors, a possible effect of un-measured variables such as C-reactive protein cannot be excluded. Third, our study participants were recruited from rural area of northeast of China, whether the re-sults can be applied to populations of different areas or races require more studies to investigate. Finally, our diagnosis based on a single biochemistry test, the cre-atinine value could be influenced by uncertain factors, thus the accuracy of eGFR could be mildly disturbed.
Conclusions
In summary, our work was the first study to reveal the relationship between CMI and reduced eGFR which was independent of prevalent CVD, medication used, and conventional cardiovascular risk factors. Thus, these data provide strong evidence of a unique, independent and economic role of CMI in a high burden of kidney disease. These findings have important implications for guiding primordial prevention and understanding the mechanisms underlying VAT-mediated renal injury.
Abbreviations
CVD: Cardiovascular disease; DBP: Diastolic blood pressure; DM: Diabetes mellitus; eGFR: Estimated glomerular filtration index; ESRD: End stage renal disease; FPG: Fasting plasma glucose; HDL-C: High density lipoprotein cholesterol; IL: Interleukin; IR: Insulin resistance; LAP: Lipid accumulation product; LCAT: Lecithin cholesterol acyl transferase; LPL: Lipoprotein lipase; NCRCHS: Northeast China Rural Cardiovascular Health Study; NO: Nitric oxide; ORs: Odds ratios; RAAS: Renin-angiotensin-aldosterone system; ROC: Receiver operating characteristic curve; SAT: Subcutaneous adipose tissue;
SBP: Systolic blood pressure; SD: Standard deviation; TG: Triglyceride; TG/ HDL-C: Triglyceride to high density lipoprotein cholesterol ratio;
TNF-α: Tumor necrosis factor alpha; VAI: Visceral adiposity index; VAT: Visceral adipose tissue; WC: Waist circumference; WHtR: Waist to height ratio
Acknowledgements
We would like to express our gratitude to all those who exert their effects in achieving this study.
Funding
This study was supported by grants from the“Thirteenth Five-Year”program funds (The National Key Research and Development Program of China, Grant #2017YFC1307600). The founder provided financial support for our survey, but do not interfere the design, operation and analysis of the study.
Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Authors’contributions
In this study, HYW and WRS did the study design, statistical analyses and results interpretation. XY, SZW and SYL participated as analyzing and resolving difficulties of analytic strategies and results discussion. Finally, YXS functioned as final reviewer who gave constructional suggestions for the interpretation of data. The corresponding author was YXS. All authors have read and approved the manuscript.
Ethics approval and consent to participate
This study was performed in compliance with the ethical principle of the Declaration of Helsinki. Written informed consents were acquired from all participants and all procedures were performed in accordance with the ethical standards. The Ethics Committee of China Medical University (Shenyang, China) approved the study protocol.
Consent for publication
Not applicable.
Competing interests
The authors declare that they have no competing interests.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Author details
1Department of Cardiology, The First Hospital of China Medical University,
155 Nanjing North Street, Heping District, Shenyang 110001, China.
2
Department of Cardiovascular Medicine, Beijing Moslem Hospital, Beijing 100054, China.3Department of Computational Medicine and Bioinformatics,
University of Michigan, 100 Washtenaw Avenue, Ann Arbor, MI 48109, USA.
4West China School of Medicine, Sichuan University, #37 Guoxue Alley,
Chengdu 610041, China.
Received: 23 July 2018 Accepted: 12 October 2018
References
1. Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet 2015;385:117–71.
2. Zhang L, Wang F, Wang L, Wang W, Liu B, Liu J, et al. Prevalence of chronic kidney disease in China: a cross-sectional survey. Lancet. 2012;379:815–22.
3. Horowitz B, Miskulin D, Zager P. Epidemiology of hypertension in CKD. Adv Chronic Kidney Dis. 2015;22:88–95.
4. Levey A, Astor B, Stevens L, Coresh J. Chronic kidney disease, diabetes, and hypertension: what's in a name? Kidney Int. 2010;78:19–22.
5. Nugent R, Fathima S, Feigl A, Chyung D. The burden of chronic kidney disease on developing nations: a 21st century challenge in global health. Nephron Clin Pract. 2011;118:c269–77.
6. Luyckx V, Tonelli M, Stanifer J. The global burden of kidney disease and the sustainable development goals. Bull World Health Organ. 2018;96:414–22D. 7. Hager M, Narla A, Tannock L. Dyslipidemia in patients with chronic kidney
disease. Rev Endocr Metab Disord. 2017;18:29–40.
8. Mallamaci F, Tripepi G. Obesity and CKD progression: hard facts on fat CKD patients. Nephrol Dial Transplant. 2013;28:iv105–iv8.
9. Hou X, Wang C, Zhang X, Zhao X, Wang Y, Li C, et al. Triglyceride levels are closely associated with mild declines in estimated glomerular filtration rates in middle-aged and elderly Chinese with normal serum lipid levels. PLoS One. 2014;9:e106778.
10. Shimizu M, Furusyo N, Mitsumoto F, Takayama K, Ura K, Hiramine S, et al. Subclinical carotid atherosclerosis and triglycerides predict the incidence of chronic kidney disease in the Japanese general population: results from the Kyushu and Okinawa population study (KOPS). Atherosclerosis. 2015;238:207–12.
11. Lee P, Chang H, Tung C, Hsu Y, Lei C, Chang H, et al. Hypertriglyceridemia: an independent risk factor of chronic kidney disease in Taiwanese adults. Am J Med Sci. 2009;338:185–9.
12. Tsuruya K, Yoshida H, Nagata M, Kitazono T, Iseki K, Iseki C, et al. Impact of the triglycerides to high-density lipoprotein cholesterol ratio on the incidence and progression of CKD: a longitudinal study in a large Japanese population. Am J Kidney Dis. 2015;66:972–83.
13. Ho C, Chen J, Chen S, Tsai Y, Weng Y, Tsao Y, et al. Relationship between TG/HDL-C ratio and metabolic syndrome risk factors with chronic kidney disease in healthy adult population. Clin Nutr. 2015;34:874–80. 14. Wen J, Chen Y, Huang Y, Lu Y, Liu X, Zhou H, et al. Association of the TG/
HDL-C and non-HDL-C/HDL-C ratios with chronic kidney disease in an adult Chinese population. Kidney Blood Press Res. 2017;42:1141–54.
15. Kramer H, Luke A, Bidani A, Cao G, Cooper R, McGee D. Obesity and prevalent and incident CKD: the hypertension detection and follow-up program. Am J Kidney Dis. 2005;46:587–94.
16. Fox C, Larson M, Leip E, Culleton B, Wilson P, Levy D. Predictors of new-onset kidney disease in a community-based population. JAMA. 2004;291: 844–50.
17. Ejerblad E, Fored C, Lindblad P, Fryzek J, McLaughlin J, Nyrén O. Obesity and risk for chronic renal failure. J Am Soc Nephrol. 2006;17:1695–702. 18. Bacopoulou F, Efthymiou V, Landis G, Rentoumis A, Chrousos G. Waist circumference, waist-to-hip ratio and waist-to-height ratio reference percentiles for abdominal obesity among Greek adolescents. BMC Pediatr. 2015;15:50.
19. He Y, Li F, Wang F, Ma X, Zhao X, Zeng Q. The association of chronic kidney disease and waist circumference and waist-to-height ratio in Chinese urban adults. Medicine (Baltimore). 2016;95:e3769.
20. Lin C-H, Chou C-Y, Lin C-C, Huang C-C, Liu C-S, Lai S-W. Waist-to-height ratio is the best index of obesity in association with chronic kidney disease. Nutrition. 2007;23:788–93.
21. Odagiri K, Mizuta I, Yamamoto M, Miyazaki Y, Watanabe H, Uehara A. Waist to height ratio is an independent predictor for the incidence of chronic kidney disease. PLoS One. 2014;9:e88873.
22. Dong Y, Wang Z, Chen Z, Wang X, Zhang L, Nie J, et al. Comparison of visceral, body fat indices and anthropometric measures in relation to chronic kidney disease among Chinese adults from a large scale cross-sectional study. BMC Nephrol. 2018;19:40.
23. Madero M, Katz R, Murphy R, Newman A, Patel K, Ix J, et al. Comparison between different measures of body fat with kidney function decline and incident CKD. Clin J Am Soc Nephrol. 2017;12:893–903.
24. Després J, Lemieux I, Bergeron J, Pibarot P, Mathieu P, Larose E, et al. Abdominal obesity and the metabolic syndrome: contribution to global cardiometabolic risk. Arterioscler Thromb Vasc Biol. 2008;28:1039–49. 25. Wakabayashi I, Daimon T. The "cardiometabolic index" as a new marker
determined by adiposity and blood lipids for discrimination of diabetes mellitus. Clin Chim Acta. 2015;438:274–8.
the risk of hypertension in Chinese population. Postgrad Med. 2018;130: 325–33.
27. Wang H, Sun Y, Li Z, Guo X, Chen S, Ye N, et al. Gender-specific contribution of cardiometabolic index and lipid accumulation product to left ventricular geometry change in general population of rural China. BMC Cardiovasc Disord. 2018;18:62.
28. Wang H, Chen Y, Guo X, Chang Y, Sun Y. Usefulness of cardiometabolic index for the estimation of ischemic stroke risk among general population in rural China. Postgrad Med. 2017;129:834–41.
29. Dursun M, Besiroglu H, Otunctemur A, Ozbek E. Association between cardiometabolic index and erectile dysfunction: a new index for predicting cardiovascular disease. Kaohsiung J Med Sci. 2016;32:620–3.
30. Li Z, Guo X, Zheng L, Yang H, Sun Y. Grim status of hypertension in rural China: results from Northeast China rural cardiovascular health study 2013. J Am Soc Hypertens. 2015;9:358–64.
31. Wang H, Li Z, Guo X, Chen Y, Chen S, Tian Y, et al. Contribution of non-traditional lipid profiles to reduced glomerular filtration rate in H-type hypertension population of rural China. Ann Med. 2018;50:249–59.
32. Chen S, Guo X, Dong S, Li Z, Sun Y. Relationship between lifestyle factors and hyperhomocysteinemia in general Chinese population: a cross-sectional study. Postgrad Med. 2017;129:216–23.
33. Kahn H. The "lipid accumulation product" performs better than the body mass index for recognizing cardiovascular risk: a population-based comparison. BMC Cardiovasc Disord. 2005;5:26.
34. Amato M, Giordano C, Galia M, Criscimanna A, Vitabile S, Midiri M, et al. Visceral adiposity index: a reliable indicator of visceral fat function associated with cardiometabolic risk. Diabetes Care. 2010;33:920–2. 35. Levey A, Stevens L, Schmid C, Zhang Y, Castro A, Feldman H, et al. A new
equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150: 604–12.
36. Stevens P, Levin A. Evaluation and management of chronic kidney disease: synopsis of the kidney disease: improving global outcomes 2012 clinical practice guideline. Ann Intern Med. 2013;158:825–30.
37. Chobanian A, Bakris G, Black H, Cushman W, Green L, Izzo J, et al. The seventh report of the joint National Committee on prevention, detection, evaluation, and treatment of high blood pressure: the JNC 7 report. JAMA. 2003;289:2560–72.
38. 2. Classification and Diagnosis of Diabetes. Diabetes Care. 2018;41:S13–27. 39. Zhang Z. Univariate description and bivariate statistical inference: the first
step delving into data. Ann Transl Med. 2016;4:91.
40. Tozawa M, Iseki K, Iseki C, Oshiro S, Ikemiya Y, Takishita S. Triglyceride, but not total cholesterol or low-density lipoprotein cholesterol levels, predict development of proteinuria. Kidney Int. 2002;62:1743–9.
41. Seravalle G, Grassi G. Obesity and hypertension. Pharmacol Res. 2017;122:1–7.
42. DeFronzo R, Ferrannini E, Groop L, Henry R, Herman W, Holst J, et al. Type 2 diabetes mellitus. Nat Rev Dis Primers. 2015;1:15019.
43. Yun HR, Kim H, Park JT, Chang TI, Yoo TH, Kang SW, et al. Am J Kidney Dis. 2018;72(3):400-10.
44. Britton K, Fox C. Ectopic fat depots and cardiovascular disease. Circulation. 2011;124:e837–41.
45. Dai D, Chang Y, Chen Y, Chen S, Yu S, Guo X, et al. Int J Environ Res Public Health. 2016;13(12).
46. Chen Y, Lai S, Tsai Y, Chang S. Visceral adiposity index as a predictor of chronic kidney disease in a relatively healthy population in Taiwan. J Ren Nutr. 2018;28:91–100.
47. Garofalo C, Borrelli S, Pacilio M, Minutolo R, Chiodini P, De Nicola L, et al. Hypertension and prehypertension and prediction of development of decreased estimated GFR in the general population: a meta-analysis of cohort studies. Am J Kidney Dis. 2016;67:89–97.
48. Tsioufis C, Kokkinos P, Macmanus C, Thomopoulos C, Faselis C, Doumas M, et al. Left ventricular hypertrophy as a determinant of renal outcome in patients with high cardiovascular risk. J Hypertens. 2010;28:2299–308. 49. Masson P, Webster A, Hong M, Turner R, Lindley R, Craig J. Chronic kidney
disease and the risk of stroke: a systematic review and meta-analysis. Nephrol Dial Transplant. 2015;30:1162–9.
50. Bellinghieri G, Santoro D, Mallamace A, Savica V. Sexual dysfunction in chronic renal failure. J Nephrol. 2008;21(Suppl 13):S113–7.
51. Naderi N, Kleine C, Park C, Hsiung J, Soohoo M, Tantisattamo E, et al. Obesity paradox in advanced kidney disease: from bedside to the bench. Prog Cardiovasc Dis. 2018;61:168–81.
52. Park J, Ahmadi SF, Streja E, Molnar MZ, Flegal KM, Gillen D, et al. Obesity paradox in end-stage kidney disease patients. Prog Cardiovasc Dis. 2014;56:415–25.
53. Mohamed-Ali V, Goodrick S, Bulmer K, Holly J, Yudkin J, Coppack S. Production of soluble tumor necrosis factor receptors by human subcutaneous adipose tissue in vivo. Am J Phys. 1999;277:E971–5. 54. Drechsler C, Wanner C. The obesity paradox and the role of inflammation.
J Am Soc Nephrol. 2016;27:1270–2.
55. KORN E. Clearing factor, a heparin-activated lipoprotein lipase. II. Substrate specificity and activation of coconut oil. J Biol Chem. 1955;215:15–26. 56. Vaziri N, Wang X, Liang K. Secondary hyperparathyroidism downregulates
lipoprotein lipase expression in chronic renal failure. Am J Phys. 1997;273: F925–30.
57. Vaziri N, Liang K. Down-regulation of tissue lipoprotein lipase expression in experimental chronic renal failure. Kidney Int. 1996;50:1928–35.
58. Vaziri N, Deng G, Liang K. Hepatic HDL receptor, SR-B1 and Apo A-I expression in chronic renal failure. Nephrol Dial Transplant. 1999;14:1462–6. 59. Vaziri N, Liang K, Parks J. Down-regulation of hepatic lecithin:cholesterol
acyltransferase gene expression in chronic renal failure. Kidney Int. 2001; 59:2192–6.
60. Zannis V, Chroni A, Krieger M. Role of apoA-I, ABCA1, LCAT, and SR-BI in the biogenesis of HDL. J Mol Med. 2006;84:276–94.
61. Ma L, Corsa B, Zhou J, Yang H, Li H, Tang Y, et al. Angiotensin type 1 receptor modulates macrophage polarization and renal injury in obesity. Am J Physiol Renal Physiol. 2011;300:F1203–113.
